How effectively does a decentralized marketplace aggregate information that is dispersed throughout the economy? We study this question in a dynamic setting where sellers have private information that is correlated with an unobservable aggregate state. We first characterize equilibria with an arbitrary finite number of informed traders. A common feature is that each seller's trading behavior provides an informative and conditionally independent signal about the aggregate state. We then ask whether the state is revealed as the number of informed traders goes to infinity. Perhaps surprisingly, the answer is no; we provide generic conditions under which information aggregation necessarily fails. In another region of the parameter space, aggregating and non-aggregating equilibria can coexist. We then explore the implications for policies meant to enhance information dissemination in markets. We argue that reporting lags ensure information aggregation while a partially revealing information policy can increase trading surplus.